Recently, the performance and functionality of digital cameras and digital movie cameras that use some solid-state image sensor such as a CCD and a CMOS (which will be simply referred to herein as an “image sensor”) have been enhanced to an astonishing degree. In particular, the size of a pixel structure for use in an image sensor has been further reduced these days thanks to rapid development of semiconductor device processing technologies, thus getting an even greater number of pixels and drivers integrated together in an image sensor. And the performance of image sensors has been further enhanced as well. Meanwhile, cameras that use a backside illumination type image sensor, which receives incoming light on its back surface side, not on its principal surface side with a wiring layer for the solid-state image sensor, have been developed just recently and their property has attracted a lot of attention these days. Nevertheless, the greater the number of pixels in an image sensor, the lower the intensity of the light falling on a single pixel and the lower the sensitivity of camera tends to be.
The sensitivity of cameras has dropped recently due to not only such a significant increase in resolution but also the use of a color-separating color filter itself. In an ordinary color camera, a subtractive color filter that uses an organic pigment as a dye is arranged to face each photosensitive cell of an image sensor. A color filter transmits one color component of incoming light to use but absorbs the other components of the light. That is why with such a color filter, the optical efficiency of a camera would decrease. Specifically, in a color camera that uses a Bayer color filter arrangement in which color filters in three colors are arranged using a combination of one red (R) pixel, two green (G) pixels and one blue (B) pixel as a fundamental unit, the R color filter transmits an R ray but absorbs G and B rays, the G color filter transmits a G ray but absorbs R and B rays, and the B color filter transmits a B ray but absorbs R and G rays. Consequently, the sum of the quantities of light that can be used by a color camera with the Bayer arrangement is approximately only one-third of the entire incoming light.
To overcome such a decreased sensitivity problem, Patent Document No. 1 discloses a technique for increasing the quantity of the light received by attaching an array of micro lenses to a photodetector section of an image sensor so that a greater percentage of the incoming light can be used. According to this technique, the incoming light is condensed onto photosensitive cells with those micro lenses, thereby substantially increasing the optical aperture ratio of the image sensor. And this technique is now used in almost all solid-state image sensors. It is true that the aperture ratio can be increased substantially by this technique but the decrease in optical efficiency by color filters still persists.
Thus, to avoid the decrease in optical efficiency and the decrease in sensitivity at the same time, Patent Document No. 2 discloses a technique for taking in as much incoming light as possible by using multilayer color filters (as dichroic mirrors) and micro lenses in combination. Such a technique uses a combination of dichroic mirrors, each of which does not absorb light but selectively transmits only a component of light falling within a particular wavelength range and reflects the rest of the light falling within the other wavelength ranges. As a result, only a required component of the incoming light falling within a particular wavelength range can be incident on each photosensing section with causing a significant loss of the incoming light.
FIG. 10 schematically illustrates a cross section of the image sensor of Patent Document No. 2 as viewed on a plane that intersects with its imaging area at right angles. This image sensor includes two condensing micro lenses 4a and 4b, which are respectively arranged on the surface and inside of the image sensor, an opaque member 20, photosensitive cells 2a, 2b, and 2c, and dichroic mirrors 17, 18 and 19, which are arranged so as to face photosensitive cells 2a, 2b and 2c, respectively. The dichroic mirror 17 has such a property as to transmit an R ray and reflect G and B rays. The dichroic mirror 18 has such a property as to reflect a G ray and transmit R and B rays. And the dichroic mirror 19 has such a property as to reflect a B ray and transmit R and G rays. The light that has impinged on the micro lens 4a has its luminous flux adjusted by the micro lens 4b, and then enters the first dichroic mirror 17, which transmits an R ray but reflects G and B rays. The light ray that has been transmitted through the first dichroic mirror 17 is then incident on the photosensitive cell 2a. On the other hand, the G and B rays that have been reflected from the first dichroic mirror 17 enter the second dichroic mirror 18 adjacent to the first dichroic mirror 17. The second dichroic mirror 18 reflects the G ray of the incoming light and transmits its B ray. The G ray that has been reflected from the second dichroic mirror 18 is incident on the photosensitive cell 2b. On the other hand, the B ray that has been transmitted through the second dichroic mirror 18 is reflected from the third dichroic mirror 19 and then incident on the photosensitive cell 2c that is located right under the dichroic mirror 19. In this manner, in the image sensor disclosed in Patent Document No. 2, the visible light that has impinged on the condensing micro lens 4a is not absorbed into color filters but their RGB components can be detected by the three photosensitive cells non-wastefully.
Meanwhile, Patent Document No. 3 discloses an image sensor that can minimize the loss of light by using a micro prism. Such an image sensor has a structure in which the incoming light is dispersed by the micro prism into red, green and blue rays to be received by three different photosensitive cells. Even when such an image sensor is used, the optical loss can also be minimized.
According to the techniques disclosed in Patent Documents Nos. 2 and 3, however, the number of photosensitive cells to provide needs to be as many as that of the dichroic mirrors to use or that of the color components to produce by splitting the incoming light. That is why to detect light rays in the three primary colors of RGB, for example, the number of photosensitive cells to provide should be tripled compared to a situation where traditional color filters are used, which is a problem.
Thus, to overcome such problems with the related art, Patent Document No. 4 discloses a technique for increasing the optical efficiency by using dichroic mirrors and reflected light, although some loss of the incoming light is involved. FIG. 11 is a partial cross-sectional view of an image sensor that adopts such a technique. As shown in FIG. 11, dichroic mirrors 22 and 23 are embedded in a light-transmitting resin 21. Specifically, the dichroic mirror 22 has such a property as to transmit a G ray and reflect R and B rays, while the dichroic mirror 23 has such a property as to transmit an R ray and reflect G and B rays.
Such a structure cannot receive a B ray at its photosensing section but can sense R and G rays entirely under the following principle. First, if an R ray impinges on the dichroic mirrors 22 and 23, the R ray is reflected from the dichroic mirror 22 but transmitted through the dichroic mirror 23. The R ray that has been reflected from the dichroic mirror 22 is also reflected from the interface between the light-transmitting resin 21 and the air, and then strikes the dichroic mirror 23. Then, the R ray is transmitted through the dichroic mirror 23 and then also transmitted through an organic dye filter 25 and a micro lens 26 that transmit the R ray. In this manner, even though only a part of the light is reflected from a metal layer 27, almost all of the R ray that has impinged on the dichroic mirrors 22 and 23 is incident on the photosensing section. On the other hand, if a G ray impinges on the dichroic mirrors 22 and 23, the G ray is transmitted through the dichroic mirror 22 but reflected from the dichroic mirror 23. The G ray that has been reflected from the dichroic mirror 23 is also totally reflected from the interface between the light-transmitting resin 21 and the air, and then strikes the dichroic mirror 22. Then, the G ray is transmitted through the dichroic mirror 22 and then also transmitted through an organic dye filter 24 and a micro lens 26 that transmit the G ray. In this manner, even though only a part of the light is reflected from the metal layer 27, almost all of the G ray that has impinged on the dichroic mirrors 22 and 23 is incident on the photosensing section almost without causing loss.
According to the technique disclosed in Patent Document No. 4, only one of the three color rays of RGB is lost but light rays of the other two colors can be received with almost no loss based on the principle described above. That is why there is no need to provide photosensing sections for all of the three colors of RGB. In this case, comparing such an image sensor to the one that does not use any dichroic mirrors but uses only organic dye filters to realize a color representation, it can be seen that the image capturing sensitivity can be doubled by this technique. This is because the optical efficiency achieved by using only organic dye filters is approximately one-third but the optical efficiency achieved by adopting the technique disclosed in Patent Document No. 4 is approximately two-thirds of the entire incoming light. Nevertheless, even if such a technique is adopted, one out of the three colors should be sacrificed.
Furthermore, Patent Document No. 5 discloses a color representation technique for improving, by providing light dispersing elements for photosensitive cells, the optical efficiency without significantly increasing the number of photosensitive cells to use. According to such a technique, each of the light dispersing elements provided for the photosensitive cells splits the incoming light into multiple light rays and makes those light rays incident on the photosensitive cells according to their wavelength ranges. In this case, each of the photosensitive cells receives combined light rays, in which multiple components falling within mutually different wavelength ranges have been superposed one upon the other, from multiple light dispersing elements. As a result, a color signal can be generated by performing a signal arithmetic operation on the photoelectrically converted signals supplied from the respective photosensitive cells. Even though this color representation technique is suitable to an image sensor with a cell pitch of approximately 1 μm, the color representation performance achieved tends to decline as the cell pitch increases.